How Were Joshua to Kings Put Together, and How Freely Did Chronicles Rewrite Them?

"Dtr. was not merely an editor but the author of a history which brought together material from highly varied traditions" — Martin Noth, 1943¹

"Overall, “Deuteronomists” and Chronicler(s) interpreted their people's history, they did not need to make it up" — Kenneth Kitchen, 2003²

Summary

Joshua, Judges, Samuel and Kings are anonymous.  None of them names its author or claims to have been written at the time it describes.  The question is not whether they were edited, since everyone who has studied them agrees they were, but when, from what, and how faithfully.

1. The final form is after 562 BC, and both sides agree. The last event in Kings is the release of the captive king Jehoiachin "in the first year" of Evil-Merodach of Babylon, which is 562 or 561 BC. Martin Noth, who named the "Deuteronomistic History" in 1943, dated its author just after that event. Kenneth Kitchen, the best-known conservative scholar of the subject, has an exilic prophet add that closing paragraph.
2. The books used older written records, and this can be shown from the text. Kings sends the reader to named sources 34 times, and the citations stop when each kingdom's court stops. Kings dates 28 accessions by the other kingdom's reign in 33 statements, and three pairs of these disagree with each other on their face, which is what copied records look like and not what one late author's arithmetic looks like. The narrator's "until this day" notes, 43 of them by Jeffrey Geoghegan's count, include some written while the temple stood and the northern kingdom was still in revolt against Judah.
3. The outside record confirms the frame. Of the 39 kings of the divided monarchy, 13 are named in inscriptions under Lawrence Mykytiuk's strict rules, plus David's dynasty. Every one is in the period and the order Kings gives. Of the eight Assyrian and Babylonian synchronisms that can be tested to the year, six fit Kings' numbers as Thiele's standard reconstruction reads them. Two do not fit without an extra assumption: Menahem's tribute in 738 and Hezekiah's "fourteenth year" in 701.
4. The models disagree about the editing, and the evidence cannot settle them. Noth had one exilic author. Frank Moore Cross had a first edition under Josiah and an exilic update. Rudolf Smend and Walter Dietrich had three exilic layers. Claus Westermann, Ernst Axel Knauf and others deny that one History ever existed. Conservatives such as Kitchen, Iain Provan and Gordon McConville have the books grow in stages from older records. All of them accept an editor after 562.
5. The editing includes at least one known mistake. Edwin Thiele, whose reconstruction of the Hebrew kings' dates conservatives rely on, concluded that four synchronisms in 2 Kings 17-18 "came into being because the final editor of Kings did not understand dual dating." Others rescue them with a co-regency. Either way, a later hand worked the numbers.
6. Chronicles copies closely where it copies, and it chooses freely what to copy. By a count of shared four-word sequences in the Hebrew, about 31 % of Chronicles' narrative verses (1 Chronicles 10 to 2 Chronicles 36) have a verse-for-verse parallel in Samuel-Kings, and those verses share about 83 % of their words with the parallel. Chronicles leaves out almost all of Saul, David's adultery and his sons' revolts (3 of 321 verses), and the stories of Elijah and Elisha (2 of 357). It adds Manasseh's repentance and long accounts of temple worship, changes "the Lord" who incited David's census to "an adversary," and differs in several numbers.
7. Some of Chronicles' differences are not changes. The Dead Sea Scroll 4QSam^a shows that the Chronicler's copy of Samuel sometimes differed from the standard Hebrew text, and Chronicles follows it. Frank Moore Cross concluded that the picture of a Chronicler "violently or willfully distorting Samuel and Kings to suit his fancy must be radically revised." The theological changes and the additions remain.

Verdict.  A conservative reader keeps real sources, real kings in the right order, a chronology that meets the Assyrian and Babylonian records almost everywhere it can be tested, and a Chronicler who used sources and did not simply invent.  The same reader gives up a Joshua or Samuel written by eyewitnesses in the form we now have, a History finished before the exile, and a Chronicles that reports Samuel-Kings neutrally.  For Joshua and Judges the outside evidence is silent:  nothing outside the Bible names any person in them, so their sources can be argued for only from the text.

Part 1: What the books show about their own making

Anonymous books that cite their sources

Joshua to Kings never names an author.  The Babylonian Talmud assigns Joshua to Joshua, Judges and Samuel to Samuel, and Kings to Jeremiah,³ but the books themselves make no such claim.  What they do say is where some of their material came from.

1. The Scroll of the Upright One. Joshua 10:13 says of the sun standing still, "The event is recorded in the Scroll of the Upright One." 2 Samuel 1:18 says David's lament over Saul "is written down in the Scroll of the Upright One."⁴ A book quoted for events about two centuries apart was a collection that existed before both passages were written into their present books. S. R. Driver, the Oxford critic, judged it "a national collection of songs celebrating the deeds of worthy Israelites," which "was not completed before the time of David, though the nucleus of the collection may obviously have been formed earlier."⁵
2. The annals of the kings. Kings cites a named source 34 times: "the scroll called the Annals of the Kings of Israel" 18 times, "the Annals of the Kings of Judah" 15 times, and "the Annals of Solomon" once.^{4 6} The Israel citations stop with Pekah, the second-last king. The Judah citations stop with Jehoiakim, before the last three kings, all of whom were deported. A writer inventing sources had no reason to stop citing them just where the royal courts that kept such records came to an end.
3. The prophets' records in Chronicles. Chronicles cites a different set: David's acts "are recorded in the Annals of Samuel the prophet, the Annals of Nathan the prophet, and the Annals of Gad the prophet," and Hezekiah's in "the vision of the prophet Isaiah son of Amoz, included in the Scroll of the Kings of Judah and Israel."⁴

None of these books survives.  The citations prove what the authors said about their method.  They do not by themselves prove the books existed.  The tests in the next sections bear on that.

The last event dates the final edition

The last paragraph of Kings has Evil-Merodach of Babylon, "in the first year of his reign," release Jehoiachin from prison and feed him at the royal table "for the rest of his life."⁴  Evil-Merodach (Amel-Marduk) came to the throne in 562 BC.  Noth:  "this concluding observation concerns an event of 562 B.C., and this gives us a definite terminus a quo for the date of Dtr.'s work."¹  The phrase "for the rest of his life" suggests Jehoiachin had died when it was written.  Kings does not mention the fall of Babylon in 539 or Cyrus's decree, which Chronicles ends with.⁴  So the final edition of Kings is most likely from between about 560 and 539.

Conservatives accept this.  Kitchen's own model has Kings written in stages by prophetic historians, with "the closing paragraph on the release of Jehoiachin ... added by an exilic prophet, ending the whole."²  The real disagreement is over how much older the material is that the final editor used.

The regnal frame: a record kept by two kingdoms

Every reign in Kings opens and closes with the same formula:  the king's age at accession, the length of his reign, his mother's name for Judah's kings, a verdict on his religion, a pointer to the annals, and his burial.  For the two centuries when Israel and Judah stood side by side, Kings also dates each accession by the other kingdom's reign:  "In the thirty-first year of Asa’s reign over Judah, Omri became king over Israel."⁴

A count of the King James text finds 33 such statements, dating 28 accessions from Abijah of Judah (1 Kings 15:1) to Hezekiah (2 Kings 18:1).⁷  They stop when Samaria falls.  After that Kings dates events only by Babylon's kings.

Three pairs of these statements disagree with each other on their face.

1. Jehoram of Israel becomes king "In the eighteenth year of King Jehoshaphat’s reign over Judah" (2 Kings 3:1) and also "In the second year of the reign of King Jehoram son of Jehoshaphat" (2 Kings 1:17).
2. Ahaziah of Judah becomes king "In the twelfth year of the reign of Israel’s King Joram" (2 Kings 8:25) and also "in the eleventh year of Joram son of Ahab" (2 Kings 9:29).
3. Hoshea kills Pekah "in the twentieth year of the reign of Jotham son of Uzziah" (2 Kings 15:30), though Jotham "reigned for sixteen years" (15:33), and Hoshea also becomes king "In the twelfth year of King Ahaz’s reign" (17:1).⁴

These disagreements are evidence about how the books were made.  An author who invented the reigns, or computed them from one master list, would not leave his own figures contradicting each other a few chapters apart.  Edwin Thiele, whose Mysterious Numbers of the Hebrew Kings is the standard defence of Kings' figures, describes the history of the problem:  "Certain scholars have decided that in the construction of a sound chronological system only the lengths of reigns are to be considered.  They have thought that the synchronisms were added at a late period and are irreconcilable with the years of reign."⁸  Thiele's own answer is that the figures come from records kept separately in Israel and Judah, each using its own way of counting a king's first year, with some reigns overlapping as co-regencies.  On that view, most of the contradictions are what two honest archives look like when an editor sets them side by side without adjusting them.  He treats one notice, 2 Kings 15:30, as a scribe's entry copied from the northern record:  it "has all the earmarks of being genuine and of appearing in the context in which it originally occurred."⁸

Thiele also found one place where the editor did adjust them, and got it wrong.  He judged the four synchronisms that tie Hezekiah to Hoshea (2 Kings 17:1;  18:1, 9, 10) to be "late and artificial":  "These synchronisms came into being because the final editor of Kings did not understand dual dating for Pekah."⁸  Later conservative chronologists, Siegfried Horn and Rodger Young, keep the four figures by giving Hezekiah a co-regency with his father from 729 BC.  That repair works, but it is an inference, since Kings never mentions such a co-regency.  Either way the chronology shows two layers:  records kept during the kingdoms, and a later editor who combined them.  The details are in Mesopotamian Absolute Chronology.

"Until this day": notes written before the end

The narrator of Joshua-Kings often explains that something is still the case "until this day."  Jeffrey Geoghegan counts 43 such notes from Deuteronomy to Kings;  a count of the Hebrew phrase, including its uses in speeches, finds 49 in Joshua-Kings alone.^{9 10}  Some of these could not have been written after the kingdom fell.

1. The poles of the ark in the temple's inner room: "They have remained there to this very day" (1 Kings 8:8). The temple burned in 586.
2. "So Israel has been in rebellion against the Davidic dynasty to this very day" (1 Kings 12:19). This assumes the northern kingdom still exists, so before 722.
3. "Ziklag has belonged to the kings of Judah until this very day" (1 Samuel 27:6). This assumes Judah still has kings, so after Rehoboam and before 586.
4. "So Edom has remained free from Judah’s control to this very day" (2 Kings 8:22).⁴

Geoghegan's conclusion is that "whenever there is any indication of when our phrase is being used it is preexilic," and that while some uses "could also derive from exilic or postexilic times, there is no instance where this is required."⁹  He thinks the notes point to Josiah's reign, though he grants that this "does not require a Josianic date for this history."⁹

Two facts limit the argument.  First, some notes do assume the north has fallen:  Israel "was deported from its land to Assyria and remains there to this very day" (2 Kings 17:23), and the settlers in Samaria still mix worship of the Lord with their idols, "their sons and grandsons ... to this very day" (2 Kings 17:41).⁴  Geoghegan dates these to Josiah's day, which is still before the exile, but they show the notes come from more than one time.  Second, Chronicles, written after the return from exile, copies five of the notes word for word from Samuel-Kings, including the one about the ark's poles still being in the temple.⁹  That temple had been destroyed for at least a century.  So a later writer could keep an "until this day" that was no longer true in his own day.  The notes therefore date the source the note was written in, not the final edition.  That is exactly the point at issue:  they show that pre-exilic written material lies inside a History finished after 562.

The editor's voice

Noth's case for one editor rested on three features that are easy to check.  At turning points the leading character gives a speech that "looks forward and backward in an attempt to interpret the course of events":¹  Joshua 1 and 23, Samuel in 1 Samuel 12, Solomon at the temple in 1 Kings 8.  Where no character can speak, the narrator does:  Judges 2:11-19 on the cycle of the judges, and 2 Kings 17:7-23 on why Samaria fell.  And every king of Judah is graded by one yardstick, Deuteronomy's law of one altar.  Six kings who otherwise "did right" are faulted with the same clause, "the high places were not eliminated."¹¹  This voice is real and it is the same voice across four books, which is why almost every model, conservative ones included, has some editor joining the books.  Whether that editor wrote under Josiah, in the exile, or both, is where the models part.

Part 2: Testing the History against the outside record

The books' own features show older records inside them.  The outside record tests whether those records were accurate.  Joshua and Judges cannot be tested this way, because no inscription names anyone in them.  Samuel and Kings can, from the ninth century on.

How many kings are named outside the Bible

Lawrence Mykytiuk, a Purdue librarian who wrote his doctoral thesis on identifying biblical persons in inscriptions, keeps the strictest published list.  He counts a person as confirmed only when an inscription from the period gives enough marks, such as name, father and title, to rule out a namesake.  His list of 53 includes these kings of Israel and Judah:¹²

That is 13 of the 39 kings of the divided monarchy, about one in three, plus David's dynasty.  The pattern of who is named is informative.

1. The named kings are the ones foreign powers had reason to name. Israel's kings appear from Omri, about 880 BC, when Assyria began campaigning west. Judah's appear a century later: Uzziah through his servants' seals, and from Ahaz on in Assyrian annals, once Assyria reached Judah. No king of either kingdom before Omri is named in an inscription, except David as the founder of a dynasty.
2. No named king is out of place. Every one of the 13 appears in the century and the order Kings gives, with the right kingdom and, where the inscription says, the right father or dynasty. Jehu, for example, appears on Shalmaneser III's Black Obelisk paying tribute in 841 BC, "Jehu (Iaua) of the house of Omri," the dynasty whose last king he killed.¹³
3. The outside texts do not copy Kings. The Assyrian annals record events Kings never mentions: Ahab's "2,000 chariots (and) 10,000 troops" at the battle of Qarqar in 853 BC¹³ and Jehu's tribute in 841. Agreement between sources that tell different stories is independent agreement.
4. Mykytiuk's rules leave some likely kings out. The Tel Dan Stele, an Aramaic victory inscription of the ninth century, names a king of Israel and the "House of David." Kitchen restores the two kings as Jehoram of Israel and Ahaziah of Judah, both killed in 841 according to 2 Kings 9, and calls the second restoration "extremely likely."² The names are broken, so Mykytiuk does not count them. The case for David's dynasty is set out in David, Solomon, and the United Monarchy and the list as a whole in Archaeological Confirmations of the Old Testament.

Mykytiuk's list also includes Shishak, the pharaoh Sheshonq I, whose raid into Judah in Rehoboam's fifth year (1 Kings 14:25) matches his own list of conquered towns on the temple wall at Karnak.¹²  That is the earliest point where Kings meets a dated foreign record.

Moab's version

The Mesha Stele, set up about 830 BC by the king of Moab, gives Moab's side of the story in 2 Kings 3.  "As for Omri, (5) king of Israel, he humbled Moab many years ... And his son followed him and he also said, 'I will humble Moab.'"  It says Israel held Medeba "in his time and half the time of his son (Ahab), forty years."¹⁴  Kings has Mesha paying tribute and rebelling "When Ahab died."⁴  The names and the order agree.  The timing does not quite:  Mesha has Moab free during the son's reign, Kings after his death.  Neither text is likely to be precise about when a slow rebellion began, but the difference is real and a defender should not hide it.

Synchronisms that can be tested to the year

The Assyrian eponym list, anchored by the solar eclipse of 15 June 763 BC, and the Babylonian chronicles allow seventeen points where Kings can be compared with a dated foreign record.  The article Mesopotamian Absolute Chronology tabulates all seventeen.  Eight are firm enough to test to the year.

So six of the eight fit Kings' numbers as Thiele's standard reconstruction reads them, and two need an extra assumption that Kings does not state.¹⁵  For a book whose critics once held its numbers to be late inventions, that is a strong result.  For a reader who expects every number to fit with no help, it is not a perfect one.

The 597 entry shows the kind of difference that remains.  The Babylonian Chronicle:  "In the seventh year, the month of Kislev, the king of Akkad mustered his troops, marched to the Hatti-land, and encamped against (i.e. besieged) the city of Judah and on the second day of the month of Adar he seized the city and captured the king."¹⁶  Kings:  "The king of Babylon, in the eighth year of his reign, took Jehoiachin prisoner."⁴  Donald Wiseman, who published the Chronicle, explains the one-year gap:  the exiles left "in the month following the capture of the city, which month also marked the commencement of Nebuchadrezzar’s eighth year."¹⁶  Two independent records agree on the king, the city, the season and the outcome, and differ by a convention of counting.

Babylon also kept the captive king.  Ration tablets from Nebuchadnezzar's palace, dated between his tenth and thirty-fifth years, list oil "to Ia-ku-ú-ki-nu, the son of the king of Ia-ku-du (i.e. Judah)" and "for the 5 sons of the king of Judah."¹⁴  Kings' last paragraph has the same man still in Babylon, still called king, and fed by the Babylonian court.⁴

Royal records in the neighbouring kingdoms

Kings' "Annals of the Kings of Israel" and "of Judah" belong to a kind of record that existed in every neighbouring state.  Assyria kept the eponym chronicle and annals of each campaign year;  Babylon kept the chronicle series that dates the fall of Jerusalem to the day.  Kitchen compares Kings' relation to its sources with the Babylonian Chronicle's relation to the fuller records behind it:  the Chronicle "is not itself the ultimate chronographic source in Babylonia (any more than was Kings in Judah/Israel), but drew upon a more extensive series of running reports."²  His conclusion on the numbers:  "Such a legacy would, most logically, derive from then-existing archives (such as the “book(s) of the annals of the kings of Judah” and “of Israel” mentioned in Kings)."²

Alan Millard, an Assyriologist at Liverpool, notes how much the outside record depends on Kings:  "No text earlier than about 850 BC names either of those kingdoms," and the first that do "would leave an insoluble puzzle were it not for the complementary biblical reports."  He is careful about the sources themselves:  "The compilers of Samuel and Kings refer to some earlier records which were at their disposal, whether or not they also drew on royal stelae written in Hebrew it is impossible to say."¹⁷

The comparison shows that such records were normal, not that Israel's existed.  No Israelite or Judahite royal annal has been found;  papyrus and leather do not survive in the hill country.  What can be said is that the details Kings gives, its reign lengths, its synchronisms and its foreign names, behave the way copied archival data behave, and not the way a later writer's guesses behave, at every point where they can be tested.

What the outside record does not show

1. Nothing before the ninth century outside David's dynasty name and Shishak's raid. The confirmed persons cluster between about 950 and 400 BC.¹² The conquest in Joshua and the judges have no outside witness at all. For those books the argument for older sources rests on the text alone: the Scroll of the Upright One, the "until this day" notes, and the archaic poetry such as Judges 5.
2. Nothing about the speeches and verdicts. The inscriptions test names, order and dates. They cannot test whether Joshua said what Joshua 23 reports, or whether God's verdicts on the kings are right. Those passages are the editor's frame on any model.
3. Nothing about when the editing happened. Accurate records can be copied by a Josianic editor, an exilic one, or both. The inscriptions show the sources were good; they do not choose among the models in Part 3.

Part 3: The models of how the History was edited

Every model below accepts the facts in Parts 1 and 2:  a final edition after 562 and older sources inside it.  They differ over how many editors there were, when they worked, and how much they changed.  Each is stated here in its makers' words.

One exilic author: Noth, 1943

Martin Noth argued that Deuteronomy through Kings is one work.  "Dtr. was not merely an editor but the author of a history which brought together material from highly varied traditions and arranged it according to a carefully conceived plan.  In general Dtr. simply reproduced the literary sources available to him."¹  He named the author for his likeness to Deuteronomy:  "we call him the "Deuteronomistic" author because his language and way of thinking closely resemble those found in the Deuteronomic Law."¹  And he dated him by the last event:  "We have no reason to put Dtr. much later than this terminus a quo."¹  On this model the author wrote in the exile to explain why the kingdoms fell, and he was a compiler who copied his sources more than he rewrote them.

A Josianic edition and an exilic update: Cross, 1973

Frank Moore Cross noticed that the History has two themes that pull against each other:  the sin of Jeroboam that doomed the north, and God's lasting promise to David that culminates in Josiah.  He argued that these "belong properly to a Josianic edition of the Deuteronomistic history," a work that "may be described as a propaganda work of the Josianic reformation and imperial program."¹⁸  After 586 an editor "retouched or overwrote the Deuteronomistic work to bring it up to date in the Exile."¹⁸  Richard Nelson, Cross's best-known follower, still holds that the History "was first composed during the reign of Josiah and then revised by the addition of limited amounts of more pessimistic material in the early exile."¹⁹  Supporters point to 1 Kings 13:2, where a prophet names "a son named Josiah" some three centuries in advance,⁴ and to the praise of Josiah as the king like no other.

Three exilic layers: Smend and Dietrich, 1971-1972

At Göttingen, Rudolf Smend and his student Walter Dietrich kept Noth's exilic date but found more than one hand.  Smend's summary:  "We have distinguished three layers:  the basic conception of the historical work (DtrH), a revision that inserts prophetic texts (DtrP), and a further one whose chief interest is the law (DtrN)."²⁰  Dietrich dated them closely:  the basic history "shortly after the conquest of Jerusalem, so probably before 580," the legal layer "shortly after the rehabilitation of Jehoiachin, so about 560," and the prophetic layer between, "between 580 and 560 BC in Judah, probably in Jerusalem."²¹  He did not think this overturned Noth:  "Noth's grand design, precisely in its single-line character, is in the main correct."²¹

Several stages, Assyrian to Persian: Römer, 2005

Thomas Römer of Lausanne takes the seventh century as "the starting point for the Deuteronomistic literary production," but holds that "it is much more likely that such a ‘history’ was not conceived before the exilic period, in an attempt by the former royal scribes to deal with the national and theological crisis of 597/587," and that "this History underwent a new redaction in the Persian period."²²  On Noth's 560 date he writes, "this view has been challenged with good reasons," citing evidence that Deuteronomistic additions kept entering the Hebrew text after the Greek translation was made.²²

No single History: Westermann, Knauf, Noll

A minority deny that one History ever existed.  Kurt Noll proposes "to replace Martin Noth's Deuteronomistic History hypothesis":  the books from Joshua to Kings are "not a deuteronomistic ideology, but a Deuteronomic debate," each book responding to Deuteronomy in its own way, so that "What we do not have, except for a few late glosses, is deuteronomism."²³  Claus Westermann (1994) and Ernst Axel Knauf (2000) argued along similar lines;  their works were not obtained for this article.  McConville summarises Westermann:  "Westermann has not actually abolished Dtr. He has simply argued that he is not a ‘historian’."  Westermann's central claim, on McConville's account, is that there was "a pre-Deuteronomistic narrative stretching from Exodus to Kings."²⁴

Conservative models

Conservative scholars do not deny editing.  They differ from the critical models mainly on how old and how reliable the sources were, and on how much the editors added.

1. Kitchen has Kings written in prophetic stages, "A first book of Kings" down to Samaria's fall, later books added, and "Finally, a generation later, the closing paragraph on the release of Jehoiachin was added by an exilic prophet, ending the whole."² This is close in shape to Cross's double edition, with more stages and earlier sources.
2. Iain Provan, V. Philips Long and Tremper Longman, in the standard evangelical Biblical History of Israel, concede the date: "The so-called Deuteronomistic History, of which Samuel—Kings would be a part, must have been completed in the Babylonian exile—whatever may have been the composition-history of its different parts."²⁵ Provan's own monograph of 1988 argued for a first edition under Hezekiah; it was not obtained for this article.
3. David Howard, in an evangelical introduction to the historical books, keeps the term only as a description: "The term thus can be used adjectivally, with no conclusions concerning authorship or date of Deuteronomy implicit in its use." He calls the Deuteronomistic History and the Hexateuch "modern critical constructs" that "cut across the present canonical division."²⁶ On Kings itself he agrees with Noth's date: "the final form of the book must date after that time, regardless of the dates of the earlier sources." Noth's ca. 550 BC, he adds, "accords well with the internal data."²⁶ He thinks the "official court records probably formed the basis for most of the information in the framework," but that the religious verdicts and even the synchronisms were the final author's, since "it is difficult to imagine official records of such hated rivals as Israel and Judah acknowledging each other in this way."²⁶
4. Gordon McConville, surveying the field for evangelical students, finds most persuasive a compromise in which "the material of the narrative existed at various stages in blocks, and that these were united into a coherent narrative by a transmission process that is lost to us," and he objects to Smend's school that there is "a serious problem with the assumption that an exilic writer created the unconditional Davidic promise, when the exile itself had made such a concept problematical."²⁴
5. Baruch Halpern, not a conservative but the leading defender of the editors as historians, argued that "some of these authors — those who wrote works recognizably historical — had authentic antiquarian intentions. They meant to furnish fair and accurate representations of Israelite antiquity."²⁷ His own model has a Hezekian history later re-edited under Josiah. He thinks the quarrel over sigla exaggerates the disagreement: "was the first redactor Hezekian or exilic? In fact, the differences are matters of detail."²⁷

What separates the models

The models disagree about the middle of the story.  None disagrees about the end:  every one, conservative or critical, has an editor working after 562.  And none claims the editors invented the kings, their order or their dates.  Nelson, defending Cross's model in 2009, admitted where the field stands:  "My objection takes us back to the impasse that has stymied us since the publication of Frank Cross's seminal article advocating a double redaction approach exactly forty years ago.  I confess I have no idea how to overcome this stalemate."²⁸

Part 4: How freely Chronicles rewrote Samuel and Kings

Chronicles retells the history from Saul's death to the fall of Jerusalem, and ends with Cyrus's decree of 539.  Its genealogy of David's line runs several generations past Zerubbabel (1 Chronicles 3).  Frank Moore Cross put the Chronicler's work "toward 400 B.C.";²⁹ H. G. M. Williamson, in the standard British commentary, judged "a date for Chronicles within the fourth century BC is most probable."³⁰  So the Chronicler wrote two centuries or more after the final edition of Kings, with Samuel-Kings in front of him.  That makes Chronicles the one place where an ancient Israelite historian can be watched at work on a source that still exists.

What a count shows

A computer comparison of the Hebrew text matched each verse of Chronicles with the verse of Samuel-Kings that shares the most four-word sequences with it, counting words by their dictionary form.³¹

Two findings come out of the table.

1. Where the Chronicler copies, he copies closely. The parallel verses share more than four-fifths of their words with Samuel-Kings. Solomon's temple prayer (2 Chronicles 6) and the account of Ahab's last battle (2 Chronicles 18) are matched almost verse for verse. A looser test, which also catches paraphrase and counts some stock phrases by accident, raises the share of parallel verses in the narrative from 31 % to about 43 %. So a little under half of Chronicles' story is Samuel-Kings, mostly word for word, and the rest is his own selection from other material.
2. What he leaves out is not random. Of the Samuel-Kings verses in each block, the strict test finds these used by Chronicles:

The pattern is plain.  The Chronicler keeps David as founder of the temple and the dynasty, and drops David's adultery, his murder of Uriah, and the revolts of his sons.  He keeps Judah's kings and drops the northern kingdom except where it touches Judah, so the Elijah and Elisha stories vanish.  Kitchen, defending the Chronicler, puts it this way:  Chronicles highlights "David’s positive achievements, for didactic purposes, but omitting his lapses (except over a census, 1Chron, 21)."²  Kitchen compares this with Babylon's Esarhaddon Chronicle, which leaves out episodes unfavourable to its king that the main Babylonian Chronicle reports.  The comparison is fair, and it concedes the point:  the Chronicler wrote a selective history with a purpose.

Hezekiah's reign shows another method.  Chronicles tells Sennacherib's invasion in 23 verses (2 Chronicles 32:1-23) against Kings' 62 (2 Kings 18:13-19:37), and in its own words, so only 6 of Kings' 95 verses match closely.  Where the Chronicler retold rather than copied, he wrote freely.

What he added

The largest additions are about worship:  the Levites' and singers' duties (1 Chronicles 15-16 and 23-27), Hezekiah's great Passover (2 Chronicles 30), and the temple's organisation.  Two additions bear directly on history.

1. Manasseh's captivity and repentance. Kings makes Manasseh the king whose sins doomed Jerusalem: "I will wipe Jerusalem clean, just as one wipes a plate on both sides."⁴ Chronicles adds that Assyrian commanders took him to Babylon with hooks and bronze chains, that "Manasseh asked the Lord his God for mercy and truly humbled himself," and that God brought him back to Jerusalem.⁴ Part of the frame is confirmed: Esarhaddon's and Ashurbanipal's inscriptions list "Manasseh, king of Judah" among their tributaries.¹² No outside text mentions his arrest. Williamson, who accepts that a historical arrest is possible, still concludes that "there can be no doubt that Manasseh's repentance has been deliberately composed by the Chronicler on the basis of 7:14," the promise that God will forgive a people who humble themselves.³⁰ Sara Japhet of the Hebrew University, a leading scholar of Chronicles, finds that "Chronicles exhibits a certain tendency to mitigate and minimize the sins of Manasseh."³²
2. Josiah's earlier reform. Kings puts Josiah's whole reform after the law scroll was found in his eighteenth year. Chronicles has him begin purging the high places in his twelfth year and find the scroll six years later.⁴ This is discussed in Deuteronomy and the Pious-Fraud Charge.

What he changed: the census

The clearest theological change is the cause of David's census.

1. 2 Samuel 24:1: "The Lord’s anger again raged against Israel, and he incited David against them, saying, “Go count Israel and Judah.”"
2. 1 Chronicles 21:1: "An adversary opposed Israel, inciting David to count how many warriors Israel had."⁴

The Hebrew of Chronicles has satan, adversary, with no article, and the same verb, "incited."  Every commentator read for this article agrees that the change is deliberate.  They disagree about what it means.

1. Williamson thinks the Chronicler, "finding his Vorlage theologically difficult at this point," borrowed the figure from Job: "it was reflection on Job in particular which led him to introduce the figure of Satan where 2 Sam. 24:1 has 'the anger of the Lord'."³⁰
2. Japhet argues that the word here is a name meaning simply "adversary" and not the later devil: "Once the name is understood as 'adversary', we no longer associate it with the later figure of Satan." She notes that Chronicles "omits the opening verse and thus the theological solution."³²
3. W. H. Bennett, writing in 1894, thought the Chronicler "must have read with something of a shock that Jehovah moved David to commit the sin of numbering Israel," and so "ascribes it to the well-known and crafty animosity of Satan against Israel."³³
4. Martin Selman, a conservative evangelical commentator, rejects the idea that the Chronicler found his source deficient: "It is clear in 2 Samuel 24:1 that God incited David because he was angry with him." But he agrees that "By changing the subject of the verb incited from 2 Samuel 24:1, it is therefore given quite a different meaning."³⁴
5. Gleason Archer, the best-known conservative harmoniser, holds that both statements are true: God allowed and Satan carried out. "Why would Satan get himself involved in this affair if God had already prompted David to commit the folly he had in mind? It was because Satan found it in his own interest to do so."³⁵

On any of these readings the Chronicler did not copy his source here.  He changed who did the inciting.  A conservative can hold, with Archer, that both writers describe one event from two sides, as Job 1-2 does.  He cannot hold that Chronicles simply reproduces Samuel.

What he changed: the numbers

Several numbers in parallel passages differ.

The verses are quoted in the footnote.⁴  The explanations offered are of three kinds, and conservatives themselves do not agree on them.

1. Copying error. Williamson allows that the tenfold charioteer figure may be due "to a desire by the Chronicler (or a predecessor of his) to magnify David's victory or to a mechanical error at some stage in the course of transmission."³⁰ Hebrew numbers written in words are easy to garble.
2. Different things counted. Archer suggests the 1,100,000 adds 300,000 reserves to Samuel's 800,000 veterans.³⁵ For the price, conservatives commonly say that Samuel gives the price of the threshing floor and Chronicles the price of the whole site, which became the temple mount.
3. Unexplained. Selman calls the usual critical reconciliation of the census "unjustified and inconsistent" and says that "the whole matter remains an unresolved issue of scribal transmission."³⁴ Bennett: "Textual corruption is always a possibility in case of numbers, but on the whole this particular change does not admit of a satisfactory explanation."³³

The honest summary is that the numbers have not been reconciled, and the leading conservative commentaries say so.

Some differences were in his copy of Samuel

The Dead Sea Scrolls changed the question.  Cave 4 at Qumran held a large Samuel scroll, 4QSam^a.  Where it survives in passages that Chronicles also tells, it often agrees with Chronicles against the standard Hebrew text of Samuel.  Cross, who led its early study:  "in sections where Chronicles overlaps with Samuel ... the text of Chronicles normally agrees with 4Q and LXX against MT."²⁹  His conclusion:  "The usual picture painted of the Chronicler violently or willfully distorting Samuel and Kings to suit his fancy must be radically revised."²⁹  Eugene Ulrich, who later edited the Qumran biblical scrolls, tested whether the scroll had instead been corrected from Chronicles and concluded that the evidence "points to the Samuel tradition, not the Chronicles tradition, as the source of the 4Q C similarity."³⁶

The census chapter gives the clearest single case.  1 Chronicles 21:16 says David "saw the angel of the Lord standing between the earth and sky with his sword drawn and in his hand, stretched out over Jerusalem," and that he and the elders fell on their faces in sackcloth.⁴  The standard Hebrew of 2 Samuel 24 has no such verse.  The scroll does:  its fragment 164 reads, with the editor's restorations in brackets, "[his eyes and saw the angel of the LORD standing between] the earth and the heavens, and his sword drawn in his hand, stretched out [over Jerusalem, and the elders fell on] their faces, covered in sackcloth."³⁷  Williamson saw a preliminary report of this fragment in 1982 and wrote that "Regrettably, this text has still not been published."³⁰  So a verse that looked like the Chronicler's own dramatic addition was in a copy of Samuel.

The scroll's limits matter as much as its evidence.

1. It does not preserve 2 Samuel 24:1, so it cannot say whether any copy of Samuel had "an adversary."
2. It does not preserve the charioteer number at 2 Samuel 10:18 or the census figures.
3. It explains some differences in wording, not the Chronicler's choices of what to omit and add. Japhet, who accepts the Qumran evidence, still concludes: "It may be that a few of these differences already appeared in the Chronicler's Vorlage, but most of them were his own work."³²

What Chronicles shows about ancient Israelite historians

Chronicles is a controlled test of how an Israelite historian handled a source, and the result cuts both ways.

1. He did not invent his material from nothing. He copied long stretches nearly word for word, sometimes from a different text of Samuel than ours, and he cites other prophetic records. Japhet, no conservative, writes that his effort to fit northern prophets into Judah's history "shows that the Chronicler was working with an existing tradition and did not 'invent facts.'"³² Weinfeld says the same of Chronicles' reports of Judah's expansion: "one nevertheless need not necessarily see in these descriptions pure invention."³⁸
2. He did reshape his material. He chose what to tell for a purpose, retold some events in his own words, changed at least one theological cause, and added speeches and episodes no earlier source has.

The inference most readers draw is that the editors of Joshua-Kings probably worked the same way on their own sources:  close copying of records where they had them, a frame and speeches of their own, and selection for a purpose.  That inference is reasonable, but it is an analogy.  The sources behind Kings are lost, so its editors cannot be watched at work the way the Chronicler can.

Part 5: What a conservative reader keeps and gives up

What the evidence supports.

1. The History drew on written sources older than its final edition: named annals whose citations end when the courts ended, a poetry collection cited twice, two kingdoms' regnal records whose disagreements were left standing, and notes written while the temple stood.
2. Those sources were good. Thirteen kings are named in foreign inscriptions in the order and period Kings gives, and the dates meet the Assyrian and Babylonian records at six of eight firm points, with the other two needing an assumption rather than a correction.
3. Chronicles is not free invention either. It copies closely, sometimes from a Samuel text older than or different from ours, and Assyrian records confirm Manasseh as a vassal.

What the evidence does not support.

1. That Joshua to Kings in its present form was written by eyewitnesses or finished before the exile. Kitchen, the most prominent conservative defender, does not claim this either.
2. That every number is as the authors wrote it. Thiele, the conservative chronologist, judged four synchronisms in 2 Kings 17-18 to be a late editor's error, and the census and chariot figures in Samuel and Chronicles still disagree.
3. That Chronicles is a neutral report. It omits David's sins and the northern prophets, adds Manasseh's repentance, and changes the cause of the census.

What the evidence cannot decide.

1. Whether the editing came in one exilic stage (Noth), a Josianic edition and an exilic update (Cross), three exilic layers (Smend and Dietrich), or a long growth of separate books (Westermann, McConville). Every model fits the data above, because the data constrain the sources and the final date, not the steps in between.
2. Whether the speeches are summaries of what was said or the editor's compositions. The outside record says nothing about them.
3. Anything about Joshua and Judges beyond what their own text shows.

A conservative doctrine of Scripture does not need an eyewitness author.  Luke says he wrote after investigating earlier accounts (Luke 1:1-4), and the Chronicler names his sources.  The question for a conservative reader is whether the compilers were faithful to their sources, not whether there were compilers.  On that question the evidence that can be checked is mostly favourable to Kings, and mixed for Chronicles' numbers.

Objections and responses

"The cited annals are a literary device.  Ancient writers invented sources to sound authoritative."  John Van Seters argued that the "archival sources" scholars find in Kings rest on a misunderstanding, since "Royal annals and commemorative inscriptions were not archival documents."³⁹  Nadav Na'aman of Tel Aviv accepts that biblical authors "did not consult archives,"⁴⁰ yet holds that the two Books of the Chronicles "must have been real sources from which the Dtr drew certain information."⁴⁰  Against the strong form of the objection:  the citations stop where real archives would stop, with Pekah in Israel and Jehoiakim in Judah, and never for the deported kings.  A writer inventing sources had no reason for that pattern.  The reign figures attached to the citations meet the Assyrian dates where they can be tested.  The objection cannot be ruled out for any one citation, but it does not explain the pattern.

"Accurate names and dates prove only that the editor had a king list, not that the stories are true."  That is correct, and this article claims no more for the inscriptions.  They test the frame.  Stories such as Elijah on Carmel or the sun standing still at Gibeon are not tested by any inscription.  What the frame shows is that the editors had good records and used them, which is evidence about their method, not proof of each story.

"The 'until this day' notes were simply copied, so they say nothing about date."  Chronicles proves that notes could be copied unchanged, and this article says so.  But that is the point:  if they were copied, they were copied from a source written while the temple stood and Israel was still in revolt against Judah.  Either way, pre-exilic written material lies behind the History.

"The synchronisms only fit because Thiele invented co-regencies to make them fit."  Part of this is fair.  Of Thiele's nine co-regencies, only two are stated in the text and the rest are inferred, as the sibling article Mesopotamian Absolute Chronology sets out.  But the Assyrian anchors at 853, 841 and 701 were not chosen by Thiele.  An invented chronology would not keep landing near dates fixed by an eclipse in 763 BC.

"Chronicles changed God to Satan, so it was willing to alter theology.  Why trust the History's editors to have done less?"  The objection is fair as far as it goes.  The Chronicler did change the theology of one verse, and his readers knew it, because Samuel was still being copied beside Chronicles.  Nothing was hidden.  The editors of Kings may well have shaped their sources as much.  That is why this article claims accurate sources and an accurate frame for Kings, not that its editors added nothing of their own.

"A conservative must hold that the Chronicler's numbers were right and Samuel's text was corrupted, or the reverse."  Not necessarily.  Selman, a conservative, leaves the census "an unresolved issue of scribal transmission."³⁴  Doctrines of inerrancy have usually applied to the original texts and allowed copying errors in later manuscripts, and the Qumran evidence shows that Samuel's text did vary.  What a conservative cannot do is treat all of the differences as copying errors, because some, like the census's cause, are clearly deliberate.

"Noth's single author has been refuted, so the whole Deuteronomistic idea is dead."  The single exilic author is disputed, and Westermann, Knauf and others doubt there was ever one History.  But the features Noth pointed to, the shared language, the speeches at turning points and one yardstick for every king, are real and have to be explained by any model.  Conservatives such as McConville explain them by later editing of books that grew separately.  No model, conservative or critical, denies that an editor worked after 562.

"If the editors wrote in the exile, their account of Israel's sins is propaganda to explain defeat."  The History does explain defeat by sin, and an exilic writer had reason to do so.  That does not make the events false.  The same History records facts against its own interest:  Judah's kings paying tribute, David's adultery, the defeat of Josiah, the twelve-year contradictions in its own numbers.  A propagandist would have smoothed those.

Issues and unanswered questions

1. The models of editing cannot be tested against each other with the evidence now available. The case for a Josianic edition rests on the History's interest in Josiah and on Geoghegan's "until this day" notes; the case for exilic layers rests on shifts in vocabulary and theme. Neither has a manuscript behind it.
2. Joshua and Judges have no outside control. The argument for old sources there rests on the Scroll of the Upright One, the "until this day" notes and archaic poetry. Weinfeld argues that Joshua 15's town list reflects Josiah's administrative districts, since En Gedi did not exist before the seventh century;³⁸ if so, at least that list was updated late.
3. The census and chariot numbers remain unreconciled. 4QSam^a does not preserve them. Three conservative commentators give three incompatible explanations.
4. Menahem's tribute in 738 and Hezekiah's fourteenth year in 701 do not fit Kings' figures without an added assumption. The co-regency of Hezekiah with Ahaz that repairs the second is inferred, not stated.
5. The Tel Dan restorations of Jehoram and Ahaziah are probable, not certain, and the editions by Biran and Naveh (1993, 1995) were not read for this article.
6. Several of the main works on Chronicles were not obtained: Japhet's 1993 commentary, Knoppers' Anchor Bible volumes (2004), McKenzie's The Chronicler's Use of the Deuteronomistic History (1985), Auld's Kings Without Privilege (1994), and Lemke's 1965 article. Auld's view, that Samuel-Kings and Chronicles both drew on an older shared story of Judah's kings, is reported here only as others describe it, and is not tested.
7. Manasseh's arrest has no outside witness. Assyrian texts confirm him as a vassal and record a Babylonian revolt in 652-648 that could explain a summons to Babylon, but no text names him in it.

Verdict

Joshua to Kings reached its present form after 562 BC.  Conservatives and critics agree on that, and the books do not claim otherwise.  The History was built from older written records:  its own citations, its unreconciled synchronisms and its pre-exilic "until this day" notes show it, and the foreign inscriptions show those records were accurate in the names, order and dates of the kings at nearly every point where they can be checked.  How many editors worked on it, and when, the evidence cannot settle.

Chronicles shows how one later historian used such records.  He copied closely where he copied, sometimes from a Samuel text different from ours, and he chose freely what to copy, what to add and, at least once, how to explain an event.  A conservative reader can keep reliable sources, a reliable frame and honest compilers.  He gives up eyewitness authorship of the present books, a History finished before the exile, perfect agreement in the numbers, and a Chronicles that merely repeats Samuel-Kings.

Sources